USPTO Patents Application 09993183

BEST AVAILABLE COPY

articles : ■ *■

Specific interference with gene function
by double-stranded RNA in early mouse
development

Florence Wianny* and Magdalena Zernicka-Goetz*f

*Welkome/CRC Institute and Department of Genetics, University of Cambridge, Tennis Court Road, Cambridge CB2 1QR, UK

fe-mail mzg@mole,bio.cam. ac.uk

V^.mX°1 do " ble *^ and ed(ds) RNA is a powerful way of interfering with gene expression in a range of organisms, but

* C0Uld successful in mammals. Here, we show that dsRNA is effective fas a
!E2f K lbltor of fun «£ on . 01 three genes in the mouse, namely maternally expressed c-mos In the oocyte and
zygotJcaNy expressed E-cadher/n or a GFPtransgene in the preimplantation embryo/The phenotypes obwrvedare *e
same as those reported for null mutants of the endogenous genes. These findings offer the op fSSriR to^dy
development and gene regulation in normal and diseased cells. y

To study early developmental events in the embryo, it is often effects of the gene knockout
deskabletobeabletoeliminateexpressionofaspecificgene. There are, nevertheless, many instances in which the existing

The most valuable information would be obtained if the "knockout' technology is extremely powerful It is however

function of the gene of interest could be disturbed in specific cells extremely laborious. It necessitates, first, making a disrupted gene

of the embryo and at defined times. In such a situation, in the segment that is suitably marked to enable the selection of homolo-

mouse, the classical techniques of gene knockout cannotbeused, gous recombination events in cultured embryonic stem cells. Such

because they eliminate gene function universally throughout the cells must then be incorporated into blastocysts and the resulting

embryo FurAermore, if a gene is repeatedly used in space and chimaeric animals used to establish pure breeding lines before
time to direct developmental processes, elimination of its role by . homozygous mutants can be obtained

gene knockout may deny an understand^ of everything but the Some of these difficulties could be overcome if a method for

first event. Even when the aim is to study the very first time in double-stranded-RNA interference (RNAi) of gene expression

development at which a gene functions, the contribution of could be developed for mammalian cells. Such an approach, first

maternal transcripts and their translation products can mask the developed in Caenorhabditis elegans', has also been drawn to be

Figure 1 MmGFP dsRNA specifically abrogates MmGFP expression In MmGFP transgenic embryos, a-c, Representative embryos out of 131 MmGFP transgenic
embryos obtaned from 11 different matings between F, females, and MmGFP transgenic males, a, Four- to six-cell embryos; b, morube; c, blastocysts. A similar pattern of
GFP express,on was obtaed after mjecbon of anbsense MmGFP RNA. <K Representative embryos out of 1 47 MmGFPtransgenic embryos that had been injected with MmGFP

!»r HZ^, f Ee " ' u " H * 1 embry ° S; °' nK,rUlae: N**"* 5 *- 8-1. Representative embryos out of 18 MmGFPtransgenic embryos that had been injected
with c-mos dsRNA at the onecell stage, g, Six-cell-stage embryos; h, morulae; I, blastocysts. Scale bars represent 20um.

70

85 © 1999 Macmlllan MagazlneSAOTRfi CELL BIOLOGY | VOL 2 1 FEBRUARY 2000 1 ceUbio.nature.,

Li

effective in other eukaryotes, including Drosophila melanogaster 1 ,
Trypanosoma bruceP> planarians 4 and plants 5 . The application of
this approach has also been demonstrated in zebrafish embryos,
but with limited success 6 . So far there has been no report that
RNAi can be used in mammals. Moreover, there are several indi-
cations of potential limitations to its function in this group of ani-
mals. Principal among these is that the accumulation of very small
amounts of dsRNA in mammalian cells following viral infection
results in the interferon response, which leads to an overall block
to translation and the onset of apoptosis 7 . Such considerations
have discouraged investigators from using RNAi in mammals.

Two factors motivated us to attempt the use of RNAi as a means
of eliminating specific gene expression in the mouse embryo. First,
we have developed approaches for microinjection of synthetic mes-
senger RNAs into both mouse oocytes and preimplantation embryos
as a means of successfully directing gene expression 8,9 . Second, we
have established a transgenic line of mice expressing a modified form
of the green fluorescent protein (MmGFP) from the ubiquitous elon-
gation factor-lot (EFlcc) promoter 10 that could provide a rapid visual
assay for the effective elimination of expression of this marker gene.
This facilitated assessment of whether RNAi could be effective in the
mouse. Here we show that it is possible to interfere with specific gene
expression in the mouse oocyte and zygote following microinjection
of the appropriate dsRNA. We show that RNAi can phenocopy the
effects of disrupting the maternal expression of the c-mos gene in the
oocyte, preventing the arrest of meiosis at metaphase n. It also inter-
feres with the zygotic expression of E-cadherin, disrupting develop-
ment of the blastocyst, as also observed in the corresponding

Figure 2 Interference wfth expression of Injected synthetic MmGFP mRNA.
a-c, Wild-type morulae injected at the onecell stage with a, MmGFP mRNA alone;
b, MmGFP mRNA together with E-cadherin dsRNA; c, MmGFP mRNA together with
MmGFP dsRNA. Scale bar represents 20 ^m.

o

articles

knockout mice. These studies show that RNAi can be effective in
mammalian cells and this fact should have substantial implications
for the analysis of gene function.

Results

dsRNA prevents expression of a GFP transgene. To determine
whether dsRNA might be used to prevent gene expression in the
mouse embryo, we developed an experimental test system using a
transgenic strain of mice that expresses MmGFP under the con-
trol of the EFlcc promoter 10 . This line offered the advantage that
GFP expression can be easily visualized in living embryos, and,
because its function is non-essential, we could monitor any non-
specific deleterious effects of dsRNA on embryonic development.
To avoid the complication of endurance of maternal gene prod-
ucts, we used heterozygous embryos in which the transgene was
paternally derived. The onset of GFP expression in these embryos
is seen by the appearance of green cells following the initiation of
zygotic transcription at the two-cell stage.

The injection of MmGFP dsRNA into the single-cell zygote pre-
vented the onset of the appearance of green fluorescence at the two-
to four-cell stages. After injection, embryos were cultured in vitro
for 3-4 days to the blastocyst stage. While uninjected embryos
expressed MmGFP in the expected manner (Fig. la-c), all embryos
injected with MmGFP dsRNA showed a markedly reduced green
fluorescence throughout this period (Fig. ld-f), with a minor pro-
portion (6.8%) showing residual weak fluorescence. The embryos
injected with MmGFP dsRNA showed normal preimplantation

Zygote Morula

Figure 3 Injection of E-cadherin dsRNA to the zygote reduces E-cadherin
expression and perturbs the development of the injected embryos.

a, Immunofluorescent staining of E-cadherin in embryos injected at the onecel! stage
with MmGFP dsRNA, and cultured for 4 days in vitro until the blastocyst stage.
b t Immunofluorescent staging of E-cadherin in embryos injected at the one-cell stage
with E<adherin dsRNA, and cultured for 4 days in vitro. Note the altered
development of these embryos. Scale bars represent 20um. c, Western blot
analysis of E-cadherin expression in zygotes, uninjected (control) morulae (collected
at the onecell stage and cultured in vitro for 3 days), morulae injected at the one-cell
stage with 2mgmr I GFP dsRNA and cultured in vitro for 3 days, and morulae injected
at the onfrcell stage with 2mg ml" 1 E-cadherin dsRNA and cultured in vitro for 3 days.
The expression of Numb protein is shown as a loading control. In each case, proteins
were extracted from 15 embryos. This experiment was repeated three times with
the same result. The reduction of signal following injection of E<adherin dsRNA was
6.5±Mold (mean±s.d.).

Table 1 Phenotypes obtained by injection of E-cadherin dsRNA Into zygotes
No. of embryos developing No. of embryos forming a
Into blastocysts after cavity after injection of E-

Experiment Injection of MmGFP dsRNA* cadherin dsRNA|

J 15/21 5/21

2 7/8 9/20

3 12/24 6/60

4 1Q/14 5/10

5 21/22 10/19

Total 66/89 (74±5%) formed 35/130 (27±4%) formed a

expanded blastocysts cavity (but did not form

expanded blastocysts). The
remainder failed to develop to

. this stage

240 uninjected zygotes were also studied. Of these, 91.6±UX (meartts.e.m.) formed

expanded blastocysts.

* MmGFP dsRNA (2 mg ml" 1 ) was injected as a control,
t 2 mg mt l Eodherin dsRNA were injected.

NATURE CELL BIOLOGY | VOL 2 1 FEBRUARY 2000 1 cellbio.naturfSS® 1999 Macmillan Magazines Ltd

71

articles

i )

Table 2 Phenotypes observed after Injection of c

-mos dsRNA Into germinal-vesicle-stage oocytes

dsRNA injected

Experiment

Number of oocytes undergoing
Spontaneous activation Fragmentation

Known null mutant phenotype

MmGFP(2mgmH)

1

0/21

0/21

NA*

2

0/22

3/22

3

1/17

0/17

4

0/13

0/13

Total

1/73 <1.4±1.4%)

3/73 (2.7±1.9%)

c-mos (2 mgml" 1 )

1

15/32

11/32

60-75% released from the metaphase-ll

arrest. HlPh rfptrrpp nf rvtnnlacmir

fragmentation 1 * 5

2

13/22

0/22

3

20/40

3/40

4

6/14

3/14

Total

53/108 (49.1 ±5%)

15/108 (13.9±3.3%)

c-mos {0.1 mg ml -1 )

1

4/17

0/17

As above

2

8/16

3/16

Total

13/33 (36 ±8%)

3/33 (6.1 ±4.2%)

development in vitro (Fig. ld-f). When transferred into pseudo-
pregnant females, they were also able to implant at the same fre-
quency as embryos derived from uninjected zygotes (40.9% and
36.1%, respectively). We compared 18 injected embryos with 22
uninjected controls at two different postimplantation stages to
determine whether they underwent normal development. Injected
embryos developed into normal gastrulating embryos at 7.0 days
post-coitum (dp.c), and were indistinguishable from control
uninjected embryos (data not shown). At 8.5 d.p.c. (three- to four-
somite stage) the injected embryos were also morphologically nor-
mal, showing that the injection of dsRNA is not toxic.

The interference with gene expression was specific, as shown by
the fact that injection of an unrelated dsRNA, corresponding to a
segment of the c-mos transcript, into MmGFP transgenic embryos
did not result in a decrease in green fluorescence (Fig. lg-i). The
injection of c-mos dsRNA did not perturb the development of the
embryos, consistent with the previous finding that the c-mos gene is
not required for normal embryonic development". Similarly, injec-
tion of dsRNA corresponding to a segment of the E-cadherin tran-
script into transgenic zygotes (59 embryos observed) did not result

Control {MmGFP dsRNA)

E-cadherin dsRNA

Figure 4 Incidence of cavity formation after Injection of E<adherin dsRNA
Into the zygote. Graphical representation of the results shown in Table 1. Dark
grey, the percentage of embryos developing into blastocysts following injection of
control MmGFP dsRNA. Light grey, the percentage of embryos forming a cavity after
injection of Eodherin dsRNA. Standard error bars are indicated.

in a decrease in green fluorescence and did not shut down protein
synthesis, although the phenotype of such embryos was abnormal
(data not shown; see also below). Transgenic zygotes injected with
antisense MmGFP RNA retained the green fluorescence at all pre-
implantation stages (37 embryos observed) (data not shown).

We attempted to detennine whether expression of MmGFP from
injected capped full-length MmGFP mRNA could be eliminated by
the co-injection of MmGFP dsRNA. We found that green fluores-
cence was gready diminished or abolished in such injected embryos
(Fig. 2c) . This was in contrast to embryos injected with sense MmGFP
mRNA, or co-injected with both sense MmGFP mRNA and the 'irrel-
evant' dsRNA for E-cadherin (Fig. 2a, b). Thus dsRNA can interfere
with the expression of both a chromosomally located gene, and of
synthetic mRNA introduced by microinjection.
Phenocopying an E-cadherin knockout We then assessed the spe-
cific developmental consequences of injecting E-cadherin dsRNA.
E-cadherin is both maternally and zygotically expressed during pre-
implantation development. Disruption of the E-cadherin gene,
using homologous recombination to remove regions of the mole-
cule esssential for adhesive function, leads to a severe preimplanta-
tion defect. These embryos can initially undergo compaction, as a
result of the presence of maternally expressed E-cadherin. How-
ever, they show a defect in cavitation and never form normal
blastocysts 12,13 .

Following injection of E-cadherin dsRNA, the phenotype was
identical to that of the null mutant embryos 12 . Thus, the embryos
initially developed normally to the compaction stage of the morula
(data not shown). However, 70% of them never formed a cavity.
The remaining 30% formed a cavity, but never developed into nor-
mal blastocysts (Table 1, Figs 3b, 4). In contrast, the majority of
uninjected embryos or control embryos injected with MmGFP
dsRNA cayitated and formed normal blastocysts (Table 1, Fig. 3a).

Analysis of E-cadherin expression t by immunostaining and
lmmunoblotting showed that the expression of E-cadherin was dra-
matically decreased after injection of E-cadherin dsRNA (Fig. 3b, c).
In contrast, no decrease in E-cadherin expression was observed in
the embryos injected with MmGFP dsRNA, in which the level of E-
cadherin expression was similar to that of the control uninjected
embryos (Fig. 3c). The level of E-cadherin at the morula stage in
embryos injected with E-cadherin dsRNA was lower than in newly
fertilized embryos before injection (Fig. 3c). This residual E-cad-
herin protein may largely reflect persistence of maternally expressed
protein whose synthesis ceases during the two-cell stage". This
residual maternal protein is present until the late blastocyst stage in

72

* © 1 999 Macmlllan MagazlneBAOTRE CELL BIOLOGY | VOL 2 1 FEBRUARY 2000 1 ceUbio.nature.com

articles

e

Figure 5 Infection of c-mos dsRNA Into Immature oocytes Inhibits c-mos
expression and causes parthenogenetic activation, a-d, Examples of
parthenogenetically activated eggs obtained after injection of c-mos dsRNA in
germinah/esicle-stage oocytes, a, Control oocyte arrested in metaphase II; b t one-
cell embryo (white arrow indicates the pronucleus); c f two-cell embryo; d, four-cell
embryo. Scale bar represents 20um. e, Western blot analysis of c-mos expression
in (left to right): oocytes arrested in metaphase II; oocytes injected at the germinal-
vesicle stage with 2mgm^ MmGfP dsRNA and cultured in vitro for 12h; and oocytes
injected at the germinah/esicle stage with 2mgmr l c-mos dsRNA and cultured m
vitro for 12h. In each case, proteins were extracted from 35 oocytes. The
expression of polo-like kinase 1 (plkl) is shown as a loading control. This experiment
was repeated three times with the same result

homozygous null embryos 12 .

We conclude that injection of E~cadherin dsRNA leads to a strik-
ing reduction in amounts of E-cadherin protein and, consequently,
a similar phenotype to that of the null mutant embryos.
dsRNA interference in the oocyte. To determine whether dsRNA
might be used to interfere with maternally expressed genes, we sought
a model gene producing a characteristic knockout phenotype. C-mos
is an essential component of cytostatic factor, which is responsible for
arresting the maturing oocyte at metaphase in the second meiotic
division. In c-mos^' mice, between 60% and 75% of oocytes do not
maintain this metaphase-II arrest and instead initiate parthenoge-
netic development 1115 . C-mos mRNA is present in fully grown imma-
ture oocytes, and its translation is initiated from maternal templates
when meiosis resumes following germinal-vesicle breakdown 16 .
Thus, injection of c-mos dsRNA would allow us to test whether
dsRNA could interfere with maternal mRNA expression.

When we injected c-mos dsRNA into oocytes, about 63% did not
maintain arrest in metaphase II (Table 2, Figs 5, 6). Of these, 78%
initiated parthenogenetic development and progressed to two- to
four-cell stage embryos (Fig. 5b-d). The remainder underwent
fragmentation. Both of these events occur at similar frequencies in
null mutant oocytes". In contrast, only 1-2% of control oocytes,
either uninjected or injected with MmGFP dsRNA, underwent
spontaneous activation (Table 2, Fig. 6). 42% of injected oocytes
failed to undergo metaphase-II arrest when we reduced the concen-
tration of injected c-mos dsRNA by 20-fold (Table 2). We con-
firmed that c-mos dsRNA interferes with c-mos expression by
immunoblot analysis carried out 12 h after the injection of germi-
nal-vesicle-stage oocytes, before the phenotypic consequences of its
loss of expression become apparent (Fig. 5e). Thus, injection of c-
tnos dsRNA into the oocyte specifically interferes with c-mos activ-
ity, mimicking the targeted deletion of c-mos by homologous
recombination. These experiments show that dsRNA is able to
block the expression of maternally provided gene products.

100n

MmGFP dsRNA omos dsRNA c~mos dsRNA
2mgmM O.lmgmM

■ Activation ■ Fragmentation

Figure 6 incidence of spontaneous activation and fragmentation after
Injection of c-mos dsRNA into the germinal-vesicle-stage oocyte. Graphical
representation of the results shown in Table 2. Oocytes were injected with the
indicated dsRNAs. The percentage of injected oocytes undergoing spontaneous
activation is shown in light grey, and the percentage undergoing fragmentation in
dark grey. Standard error bars are indicated.

Discussion

We have shown that dsRNA can be used as a specific inhibitor of
gene activity in the mouse oocyte and preimplantation embryo. We
showed the specificity of the procedure by individually inhibiting
the expression of three different genes: c-mos in the oocyte, and E-
cadherin or a GFP transgene in the early embryo. In the cases of the
two endogenous mouse genes, this results in phenotypes compara-
ble to those of null mutants. Our experiments aimed at preventing
expression of the GFP transgene indicate that RNAi per se does not
affect the normal course of development.

Thus it appears that the concerns that RNAi might not work in
the mouse may have been raised prematurely (reviewed in ref. 17).
Concern has been expressed that the protocols used for invertebrate
and plant systems are unlikely to be effective in mammals, because
accumulation of dsRNA in mammalian cells can result in a general
blockage of protein synthesis. The presence of extremely low con-
centrations of dsRNA in viral infections triggers the interferon
response 18 , part of which is the activation of a dsRNA-responsive
protein kinase (PKR) 19 . This enzyme phosphorylates and inacti-
vates translation factor EIF2cc in response to dsRNA. The conse-
quence is a global suppression of translation, which in turn triggers
apoptosis. However, we have shown here that the injection of a
dsRNA is specific to the corresponding gene; it does not cause a
general translational arrest, because embryos continue to develop
and we see no signs of cell death. It is possible that the early mouse
embryo is incapable of an interferon response, and that there may
still be difficulties in using RNAi at later stages. However, the inter-
feron response normally occurs in response to viral infection and is
usually induced experimentally using synthetic double-stranded
ribonucleotide homopolymers. 'Natural' dsRNA may be less effec-
tive at inducing PKR, and the degree of induction could vary
between cell types, in which case RNAi would be effective.

It has been suggested in other systems that genetic interference
from injected sense or antisense RNA is actually mediated by
dsRNA present at a low level in all in vitro RNA syntheses because
of the nonspecific activity of RNA polymerases 1 . Antisense RNA has
been used as a means of reducing gene expression in the embryos of
a number of species. Although it has had considerable success in
Drosophila, it has been disappointing in Xenopus, zebrafish and

NATURE CELLBIOLOGY | VOL 2 1 FEBRUARY 2000 1 cellbio.naturtS® 1999 Macmlllan Magazines Ltd

73

articles

i

i )

mouse embryos. In Xenopus, the limitations in using the antisense
approach were thought to be due to a prominent RNA-melting
activity 20 - 21 , exerted by the dsRNA-specific adenosine deaminase
(dsRAD), and which itself argues against the likelihood of success
for RNAi. However, although dsRAD has the potential to lead to the
instability of injected dsRNA and thereby might be expected to
reduce the efficacy of the approach, others have postulated that
dsRNA modifed by this enzyme might actually mediate RNAi
through the targeted degradation of endogenous RNA 17,22 . In the
mouse embryo, the use of antisense RNA has had inconsistent and
limited success in reducing gene expression, possibly because of the
instability of RNA, particularly between the two-four-cell stages 23 .
It has been recently reported that dsRNA might be more effective
than antisense RNA in inhibiting gene expression in zebrafish
embryos 6 . However, in contrast to our experiments in the mouse,
these initial experiments with zebrafish embryos indicated only
partial success of dsRNA interference.

Two of our experiments support the hypothesis that RNAi acts
in the mouse by either inducing degradation of the targeted RNA or
inhibiting its translation. First, we showed that injection of MmGFP
dsRNA inhibits the expression of co-injected sense MmGFP
mRNA. Second, we injected dsRNA against c-mos into oocytes
before the germinal vesicle breaks down, the stage when c-mos
mRNA has accumulated but has not yet been translated. C-mos is
translated when the germinal vesicle breaks down, to arrest oocytes
in metaphase of the second meiotic division 16 - 24 . We found that c-
mos dsRNA prevents c-mos function: oocytes proceed through
metaphase II and undergo parthenogenetic activation. In each case,
the effects of RNAi persist for sufficient time to phenocopy the loss
of gene function. As interfering with the expression of MmGFP is of
no consequence to the embryo, this allows us to determine how
long the RNAi effect persists. We found that although green fluo-
rescence was absent or greatly reduced by MmGFP dsRNA in trans-
genic blastocysts injected as zygotes, fluorescence did return in
embryos at 6.5 days postimplantation. During this time period,
some 10-20 cells of the inner cell mass undergo a 100-fold increase
in cell number, corresponding to a 40-50-fold increase in cell mass.
This is coincident with a parallel increase in the expression of the
transgene in uninjected embryos.

As the effects of dsRNA-mediated inhibition of gene expres-
sion persist for more than six rounds of cell division, RNAi offers
new opportunities to study loss-of-function phenotypes in spe-
cific cells and at specific stages of development of the early mouse
embryo. It should be possible to study the loss of function of not
only any single gene, but also combinations of genes, either fam-
ily members that may have redundant functions, or several mem-
bers of a regulatory pathway. Moreover, the use of RNAi can be
extended to evaluate the loss of function of particular genes dur-
ing oocyte maturation. This also provides a means of eliminating
the expression of maternally provided RNA to study maternal
effects of genes that show lethality in homozygous mutants We
anticipate that it should be equally effective in other mammals,
including both domestic animals and humans, in which it is dif-
ficult or impossible (because unethical) to create loss-of-func-
tion mutants and perform standard in vivo mutational analysis
If the approaches that we describe can be extended to the adult
organism, they will have considerable therapeutic power in
inhibiting gene activity in several types of disease. At the
moment, in addition to allowing the analysis of genes that regu-
late development, elimination of gene function by RNAi in the
mouse oocyte and preimplantation embryo should find wide-
spread application in the study of genes that regulate all basic cel-
lular processes, such as cell-cell interactions, intracellular
trafficking and the cell-division cycle. □

Methods

Collection and culture of oocytes and embryos.

Immature oocyte* arrested at prophase 1 of meiosis were collected from ovaries of 4^-weeknold F (CBA
x C57/B1) mkeinFHM medium (Speciality Media Inc. Lavalette. NI) supplemented with bovine 'serum

albumin (BSA) (4 mg ml- 1 ).

F, female mice were superovulated by intraperitoneal injections of pregnant mare's serum
gonadotroph* (PMSG, 5 international units (Lu.)) and human chorionic gonadotrophs (hCG, 5 L u.)
46-52 h "P* 1 ** Fertilized one-cell embryos were obtained from mated females 20-24 h after hCG injection.

RNA synthesis and microinjections.

The templates used for RNA synthesis were linearized plasmids. Full-length MmGFP complementary
DNA (714 base pairs (bp)) was doned into the T7TS plasmid. A KpnVHmdlU fragment of c-mos cDNA
(550 bp) was cloned into Bluescript pSK. A cDNA fragment corresponding to exons 4-8 of E-cadhenn
(580 bp) was cloned into Bluescript pKS. RNAs were synthesized using the T3 or T7 polymerase, using
the Megascripts kit ( Ambion). DNA templates were removed with DNase treatment. The RNA products
were extracted with phenol/chloroform, and ethanol-precipitated. .

To anneal sense and antisense RNAs, equimolar quantities of sense and antisense RNAs were mixed
intheannealmgbuffer (10mMTn^pH7^ 0.1^

for 10mm at 68 «C and incubated at 37°C for 3-4h. To avoid the presence of contaminating single-
stranded RNA in the dsRNA samples, the preparations were treated with 2 u g mT RNaseTl
(Calbiochem) and 1 ugmT' RNase A (Sigma) for 30m in at 37 °C The dsRNAs were then treated with
140ug ml" 1 proteinase K (Sigma), phenol/chloroform-extracted and ethanol-precipitated. Formation of
dsRNA was confirmed by migration on an agarose get: for each dsRNA, the mobility on the gel was
shifted compared to the single-stranded RNA*. For comparison of antisense and double-stranded RNAs,
equal masses of RNA were injected.

RNAs were diluted in water, to a final concentration of 2-4mgmt-'. The range of effective
concentrations is best illustrated by the c-mos RNAi experiment (Table 2) because of the sensitivity of
this biological phenotype. The mRNAs were mtcroinjected into the cytoplasm of the oocytes or embryos,
using a constant flow system (Transjector, Eppendorf) as described Each oocyte or embryo was injected
with ~10pl dsRNA. Improved penetrance was achieved by using negative capacitance. After
microinjection, oocytes and embryos were cultured in KSOM (Speciality Media Inc.) medium
supplemented with 4mgmT< BSA, at 37 °C in a 5% CO, atmosphere. MmGFP transgenic embryos were
observed by confocal microscopy (Biorad 1024 scanning head on a Nikon Eclipse 800 microscope). Some
blastocysts derived from uninjected zygotes or zygotes injected with MmGFP dsRNA were transferred
into the uteri of pseudopregnant mothers that had been mated 2.5 days earlier with vasectomized males.
Embryos were recovered either at embryonic day (E) 7.0 or £8.5, counting noon of the plug day of the
pseudopregnant recipient as E0.S. They were observed by confocal microscopy as described".

Immunoblot and immunostafning analysis.

For immunoblot analysis, samples were subjected to SDS-PAGE and proteins were transferred to a
hybond nitrocellulose membrane (Amersham). Membranes were preincubated in TBST buffer (20 mM
Tris-HCL PH8.2, 150mM Nad, 0.1 % Tween-20) containing 5% (w/v) non-fat dried milk overnight, to
block nonspecific binding of antibodies. They were then incubated with the anti-E-cadhcrin antibody
(DECMA-1), the anti-mos antibody (SantaCruz Biotechnology), the anti-Numb antibody (provided by
R. Pedersen), or the anti-plkl antibody* for 1 h, washed in TBST, incubated with the peroxidase-
conjugated secondary antibody (SantaCruz Biotechnology) for 1 h, and washed again in TBST. The
antibodies were diluted in TBST containing 5% (w(v) non-fat dried milk. The secondary antibody was
detected by enhanced chemuurainescence (Amersham). The decrease in E-cadherin expression was
quantified by comparing the optical density of the bands obtained in each western blot analysis, on a
Macintosh computer using the public-domain NIH Image program.

For whole-mount immunofluorescence with anti-E-cadherin antibody, embryos were fixed in 2%
paraformaldehyde for 20min at room temperature, followed by permeabilization with 0.1% Triton X-
100 for lOmin. After preincubation in 2% BSA in PBS for 30 min, embryos were incubated with the anti-
E-cadherin antibody for 1 h at 37 "C, and with a Texas-red-conjugated goat anti-rat antibody (Jackson
ImmunoResearch) for lhat 37«G Embryos were observed using the Biorad 1024 laser scanning confocal
microscope.

WUSHW »^l£^ lSED 25 QCT0B£R im ACCEPT£D 17 DECEMBER .999;

1. Fire, A. et al Potent and specific genetic interference by double-stranded RNA in Caenorhabditis
elegans. Nature 391, 806-811 (1998).

2. Kennerdell, J. R. & Carthew, R. W. Use of dsRNA-mediated genetic interference to demonstrate that
frizzled and frizzled 2 act in the wingless pathway. Cell 95. 1017-1026 (1998).

3. Ngo, H., Tschudi, C, Gull, K. & Ullu, E. Double-stranded RNA induces mRNA degradation in
Trypanosoma brucei. Proc Natl Acad. ScL USA 95, 14687-14692 (1998).

4. Sanchez Alvarado, A. fit Newmark, P. A. Double-stranded RNA specifically disrupts gene expression
during planarian regeneration. Proc Natl Acad. Sci. USA 96, 5049-5054 (1999).

5. Waterhouse, P. M., Graham, M. W. * Wang, M. B, Virus resistance and gene silendng in plants can
be induced by simultaneous expression of sense and antisense RNA. Proc Natl Acad, Scu USA 95
13959-13964 (1998).

6. Wargelius, A., Eflingsen. S. 8c Fjose, A. Double-stranded RNA induces specific developmental defects
in zebrafish embryos. Biochcm. Biophys. Res. Commun. 263, 156-161 (1999).

7. Lee, S. B. «c Esteban, M. The interferon-induced double-stranded RNA-activated protein kinase
induces apoptosis. Virology 199, 491-496 ( 1994).

8. Zernicka-Goetz, M.etaL Following cell fate in the living mouse embryo. Development 124, 1 133-
1137(1997).

9. Wianny, F„ Tavares, A.. Evans, M. J., Glover, D. M. & Zernicka-Goetz, M. Mouse polo-like kinase 1
associates with the acentriolar spindle poles, meiotic chromosomes and spindle midzone during
oocyte maturation. Chrcmosoma 107, 430-439 (1998).

10. Zernicka-Goetz, M. in CdlUneage and Fate Determination (ed. Moody, S. A.) 521-527 (Academic
San Diego, 1999).

1 1. Cotledge, W. H., Carlton, M. B., Udy, G. B. & Evans, M. J. Disruption of c-mos causes
parthenogenetic development of unfertilized mouse eggs. Nature 370, 65-68 (1994).

12. Larue, L, Ohsugi, M., Hirchenhain, J. & Kemler, R. E-cadherin null mutant embryos fail to form a
trophectoderm epithelium. Proc Natl Acad. ScL USA 91, 8263-S267 ( 1994).

13. Riethmacher.D, Brinkmann, V. 8c Birdimeier, C A targeted mutation in the mouse E-cadherin gene
results in defective preimplantation development Proc Natl Acad Sci. USA 92, 855-859 (1995).

74

3© 1999 Macmlllan Magazlnc&A3lPRE CELL BIOLOGY | VOL2 1 FEBRUARY 2000 1 ce^bio. nature.com

{ )

articles

14. Sefton, M., Johnson, M. H. St Clayton, L. Synthesis and phosphorylation of uvorooruUn during
mouse early development. Development 1 15, 3 1 3-31 8 ( 1 992).

15. Hashimoto, N.ctaL Parthenogenetic activation of oocytes in c-roos-deficient mice. Nature 370, 68-
71 (1994).

16. Verlhac, M. H. et aL Mos is required for MAP kinase activation and is involved in microtubule
organization during meiotic maturation in the mouse. Development 122, 815-822

(1996).

17. Fire, A. RNA-triggered gene silencing. Trends Genet. 15, 358-363 (1999).

18. Marcus, P. 1. Interferon induction by viruses: one molecule of dsRNA as the threshold for interferon
induction. Interferon 5, 115-180 (1983).

19. Clemens, M. J. PKR — a protein kinase regulated by double-stranded RNA. Int. ]. Btochem. Cell BioL
29,945-949(1997).

20. Bass, B. L fit Weintraub, H. A developmen tally regulated activity that unwinds RNA duplexes. Cell
48,607-613(1987).

21. Rebagliati, M. R. 8t Melton, D. A. Antisense RNA injections in fertilized frog eggs reveal an RNA
duplex unwinding activity. Cell 48, 599-605 (1987).

22. Sharp, P. A. RNAi and double-strand RNA. Genes Dev 13, 139-141 (1999).

23. Bevilacqua, A., Erickson, R. P.fic Hieber, V. Antisense RNA inhibits endogenous gene expression in

mouse preimplantation embryos: lack of double-stranded RNA 'melting* activity. Proc Natl Acad.
Sci USA 85, 831-635 (1988).

24. Paules, R. Buccione, R-, Moachd, R. C, Vande Woude, G. F. fir Eppig, J. J. Mouse Mos
protooncogene product is present and functions during oogenesis. Proc Natl Acad. ScL USA 86,
5395-5399 (1989).

25. Weber, R. J., Pedersen, R. A, Wianny, F, Evans, M. ;. & Zerakka-Goete, M. Polarity of the mouse
embryo is anticipated before implantation. Development 126, 5591-5598 (1999).

ACKNOWLEDGEMENTS

We thank D. Glover for suggestions and support throughout the course of this work; J. Gurdon for the
initial idea of injecting mRNA into embryos; A. McLaren for her suggestion of the use of E-cadherin as
a marker; L. Larue for E-cadherin plasmids and antibody; B. Colledge for the c-mos plasmid; P.-Y.
Bourillot for subdoning the c-mos construct and for helpful discussions; R. Weber for supportive
discussions and help in initiating microinjection experiments; M. Evans for his enthusiasm and support
with the initiation of this work; and B. Tom for the help with the statistical analysis. This work was
supported by a CRC Grant to M.Z.-G., D. Glover and M. Evans. M.Z.-G. is a Senior Research Fellow of
the Lister Institute for Preventive Medicine, and a Stanley Elmore Fellow of Sydney Sussex College.
Correspondence and requests for materials should be addressed to M.Z.-G.

NATURE CELL BIOLOGY | VOL 2 1 FEBRUARY 2000 1 ceUbio.naturlft® 1999 Macmlllan Magazines Ltd

75

This Page is Inserted by IFW Indexing and Scanning
Operations and is not part of the Official Record

Defective images within this document are accurate representations of the original
documents submitted by the applicant.

Defects in the images include but are not limited to the items checked:

□ BLACK BORDERS

□ IMAGE CUT OFF AT TOP, BOTTOM OR SIDES

□ BLURRED OR ILLEGIBLE TEXT OR DRAWING

□ SKEWED/SLANTED IMAGES

□ COLOR OR BLACK AND WHITE PHOTOGRAPHS

□ GRAY SCALE DOCUMENTS

□ LINES OR MARKS ON ORIGINAL DOCUMENT

□ REFERENCE(S) OR EXHIBIT(S) SUBMITTED ARE POOR QUALITY

□ OTHER: _____

IMAGES ARE BEST AVAILABLE COPY.
As rescanning these documents will not correct the image
problems checked, please do not report these problems to
the IFW Image Problem Mailbox.

BEST AVAILABLE IMAGES

FADED TEXT OR DRAWING